Hidden Intent Inference on MISID benchmark leaderboard · SOTA2 Research

Hidden Intent Inference on MISID

53.27Hidden Intent Inference ScoreClaude-Sonnet-4.5

Evaluation Results

Method	Links
Claude-Sonnet-4.52026.04		53.27
GPT-5.12026.04		52.83
Gemini-3-flash2026.04		49.52
DeepSeek-V32026.04		48.02
GPT-4o2026.04		46.09
DeepSeek-R12026.04		45.59
DeepSeek-V32026.04		44.23
Claude-Sonnet-4.52026.04		43.35
GLM-4.72026.04		43.04
GPT-4o2026.04		41.23
Grok-4-Fast2026.04		40.85
Grok-4-Fast2026.04		40.01
GPT-4o2026.04		39.77
Qwen3.5-Plus2026.04		38.84
GPT-5.12026.04		36.46
DeepSeek-R12026.04		36.35
Grok-4-Fast2026.04		32.69
GLM-4.72026.04		31.96
Qwen3-Max2026.04		30.44
MiniMax-M2.12026.04		29.72
Qwen3-VL-Flash2026.04		28.27
Gemini-2.5-flash2026.04		26.85
Qwen3-VL-Max2026.04		23.42
Qwen3-VL-Plus2026.04		23.27
Qwen3.5-Plus2026.04		23.27
Qwen3-Max2026.04		22.85
Qwen3.5-Plus2026.04		22.12
Qwen3-VL-235B-A22B2026.04		21.69
MiniMax-M2.12026.04		21.69